diy solar

diy solar

EDITED TITLE: Why not common Neutral? What is dangerous about it?

Zwy said:
The issue is whether or not the bonding relay in only one inverter
Click to expand...
You have to remove all bonding screws.
Zwy said:
What size of N would be required with 6 inverters?
Click to expand...
It should be sized for the current it will be carrying.
Neutral sizing is always larger or equal to ground conductor.
Zwy said:
3 pole neutral switching where the system is SDS works too.
Click to expand...
This removes the grid input from the AIO.
Which is what causes the issue in the first place.
Zwy said:
These are sold as off grid inverters,basically standalone, thus the classification requires N-G bonding to be at the inverter.
Click to expand...
Correct, if installed as an off grid stand alone system.
But, that's not how they are being installed. Which is why this issue keeps coming up.
This issue only arises from using the grid input.
 
timselectric said:
You have to remove all bonding screws.
Click to expand...

That's the problem with screw removal, it seems like an all or nothing which brings up ground fault detection and how current will return to source.

timselectric said:
It should be sized for the current it will be carrying.
Neutral sizing is always larger or equal to ground conductor.
Click to expand...

From what I see, it has to be sized according to all current carrying conductors max current.

timselectric said:
This removes the grid input from the AIO.
Which is what causes the issue in the first place.
Click to expand...
timselectric said:
Correct, if installed as an off grid stand alone system.
But, that's not how they are being installed. Which is why this issue keeps coming up.
This issue only arises from using the grid input.
Click to expand...
No, the 3 pole is placed between the inverters and load panel. This enables return to grid power not thru the AIO. As the AIO has an internal neutral switching transfer switch, the transfer switch should isolate the system as an SDS if the only source for the load panel is thru the inverter both in pass thru and inverter mode.

The issue some members think is a problem is there is the possibility of objectionable current on G between the stacked inverters. I'm more of this line of thinking. https://diysolarforum.com/threads/s...off-grid-inverter-6500ex-48.40478/post-517412
 
FilterGuy said:
If the incoming neutral is actually hot and the incoming hot is actually neutral, the system is already very dangerous. There is a very good chance the vehicle chassis is hot even if there are no inverters installed.

Furthermore, if the inverters are installed with a common neutral and no bonding screws, nothing additional would happen. The inverter can't tell a hot-neutral swap unless they measure voltage between neutral and ground.... and we know they don't based on how they behave. The inverter will happily take in the swapped input and pass it right through.

If the inverters are installed with the bonding screws there would be a short if the inverter engages the bonding relay before it disengages the hot and neutral relays. If this were to happen the breaker at the monument should pop.... If the inverter on the monument doesn't pop, all hell will break loose.....and that is without a common neutral arrangement.
Click to expand...
The breaker should pop, but it is too slow to prevent damage.
I saw it elsewhere, copper vaporized before breaker popped. But it did pop and prevented a fire.
BTW: One RV Pedestal had Neutral and Ground swapped. No problem for me and they fixed it next day.
 
My comment on why you should not have a "common neutral" on an inverter input and output. This is the way I visualize it:

1. Electricity always wants to get back to its source.

2. Ground conductor (green) should never have any current under normal conditions.

3. Purpose of ground conductor is to safely carry fault currents back to the source for two reasons (lower touch voltage on metal equipment AND to allow an immediate high current to flow to open a breaker (fuse) and clear the fault.)

4. The neutral-ground connection should be located at the currently operating source (N-G at service entrance, N-G on genset, or N-G on inverter IN INVERT MODE). The point is that the N-G connection needs to be where the electricity wants to get back to.

5. There can be no other N-G connection other than at the source or you have just created a parallel path for normal load current. If you have a N-G connection in inverter in pass through and and N-G connection at the grid service entry (or genset) than neutral current heading back to the source now has a parallel path and you will get current in the ground conductor which should never happen.

6. The current moving in the ground conductor (connected to metal cases and etc) now causes a voltage potential difference at different points on the system and a voltage potential relative to earth. A voltage potential relative to earth is a shock hazard.

And this is why Victron (which I am most familiar with) which is meant to transfer and/or invert include the N-G bond switching. When you change the source (grid vs. invert) you must change where the N-G bond is physically made.
 
Last edited:
FilterGuy said:
In NEC parlance, the Grounded Conductor is what we call neutral and it is NOT a Grounded Conductor (neutral) if it is not connected to the ground. So, yes, a neutral not only can be shorted to ground but by definition, it IS shorted to ground. If the short is not at the system grounding jumper, it becomes a 2nd Neutral-Ground bond. As @timselectric says, a GFCI upstream from the short will trip, but other than that, there is very little to indicate the problem unless you are measuring current on the ground and neutrals. (Too much current on the ground or too little current on the neutral between the N-G bonds will indicate a problem)
Click to expand...
I am trying to say that not only can we ground the neutral on purpose, it can also happen by accident. When the phase happens to short to the ground the fuse will trip, when it happens to the neutral, we end up with a ground loop. It is the ground loop which is a concern, my point is that we are trying to solve a problem that has existed since the day we decided to ground the neutral and before there were practical means to solve the issue. (assuming there is a practical solution possible)
 
l00semarble said:
The neutral-ground connection should be located at the currently operating source (N-G at service entrance, N-G on genset, or N-G on inverter IN INVERT MODE). The point is that the N-G connection needs to be where the electricity wants to get back to.
Click to expand...
It can be, but isn't required to be.
A power system is required to only have one N/G bond. And it is required to be at or before the first means of disconnect of the utility source.
This serves as the N/G bond for all sources for a system.
The only thing that changes this. Is if an auxiliary source is installed as a separately derived system. Meaning that you chose to not use a common neutral. (Which is not industry standard)
But if you choose to install it as a separately derived system. Then, you are required to add a N/G bond for the auxiliary power source.
 
timselectric said:
GFCI
Click to expand...
I could be wrong, but I think a GCFI would not work because it would require the short to occur in two place (one before the GCFI circuit and one after) It might be possible to measure a potential difference between an unbounded ground and phase or neutral. If so, it would indicate a short, which would require some one to look at it. The problem I foresee is that many people will just ignore those warnings because "everything is working fine".
 
timselectric said:
It can be, but isn't required to be.
A power system is required to only have one N/G bond. And it is required to be at or before the first means of disconnect of the utility source.
This serves as the N/G bond for all sources for a system.
The only thing that changes this. Is if an auxiliary source is installed as a separately derived system. Meaning that you chose to not use a common neutral. (Which is not industry standard)
But if you choose to install it as a separately derived system. Then, you are required to add a N/G bond for the auxiliary power source.
Click to expand...
But isn’t an inverter going to always be considered a “separately derived” source?

In the marine world we usually refer to it as a “newly derived” source. I’m not sure if this is just a nomenclature difference or if there is some technical difference.
 
l00semarble said:
4. The neutral-ground connection should be located at the currently operating source (N-G at service entrance, N-G on genset, or N-G on inverter IN INVERT MODE). The point is that the N-G connection needs to be where the electricity wants to get back to.
Click to expand...
This is where our views start to diverge.

Almost all small gensets sold in the US do NOT have an NG ground. The working assumption is that if they are plugged into a house circuit, the neutral won't be switched off and the house circuit's main bonding jumper will provide the NG bond. Unfortunately, this means a dynamic bond should be added to the genset when it is used as the AC input of a typical mobile installation that does dynamic bonding.

Inverters are a mixed bag. With the exception of Victron all of the higher-end inverters don't do dynamic bonding and have the input and output neutral tied together internally. (They have a common neutral). Almost all of the value-priced inverters do dynamic bonding and appear to disallow common neutral configurations.

Note: Unlike the value-priced inverters, Victron makes accommodations for common neutral if that is the way you want to wire it.

In my personal opinion, dynamic bonding is best left to mobile installs and stationary installs should use common neutral and no dynamic bonding.
The common neutral with a common N-G bond is much simpler and has a much higher reliability. (No relays to go bad and no controlling logic to go bad).

l00semarble said:
5. There can be no other N-G connection other than at the source or you have just created a parallel path for normal load current.
Click to expand...
I absolutely agree there should only be one NG bond. But it can be a single NG bond that services several sources as long as the neutral is not switched.
 
The problem the manufacturers are trying to address is loss of the neutral-ground bond upstream of their equipment. It is easier and cheaper (especially when talking about different international neutral bonding schemes) to consistently require a dedicated neutral (and not use an incoming neutral).
 
LeoThomson said:
I could be wrong, but I think a GCFI would not work because it would require the short to occur in two place (one before the GCFI circuit and one after)
Click to expand...
No, the GFCI does not need an NG bond before it to detect the problem. However, the problem does have to be downstream of the GFCI.
 
LeoThomson said:
I could be wrong, but I think a GCFI would not work because it would require the short to occur in two place (one before the GCFI circuit and one after) It might be possible to measure a potential difference between an unbounded ground and phase or neutral. If so, it would indicate a short, which would require some one to look at it. The problem I foresee is that many people will just ignore those warnings because "everything is working fine".
Click to expand...
A GFCI monitors current (to the mili-amp) on both current carrying conductors. If there's any deviation it opens the circuit.
In fact, a GFCI is the only code compliant way to install a three prong outlet (hot,neutral, and ground) when there's no ground available.
 
l00semarble said:
But isn’t an inverter going to always be considered a “separately derived” source?
Click to expand...
Only if it has a separate neutral. That does get into some of the subtleties though of why common vs dedicated neutral becomes a religious discussion.

The bottom line is that when they design for a dedicated neutral much of their electronics and design can be shared between different markets, which makes them cheaper to build. You can ostensibly use the same inverter on 120/240, 220V 1P, 220/380V 3P, 220/"380"V 1P, on-grid and off-grid.
 
Shimmy said:
The problem the manufacturers are trying to address is loss of the neutral-ground bond upstream of their equipment. It is easier and cheaper (especially when talking about different international neutral bonding schemes) to consistently require a dedicated neutral (and not use an incoming neutral).
Click to expand...
Losing the upstream N-G bond is what dynamic bonding is there to address.... but I believe that is primarily needed for the mobile market where shore power with it's N-G bond comes and goes.

You may also be right, about international differences, but my impression is that inverters intended for non-us markets are also a mixed bag with value-added inverters doing bonding and the higher-end inverters having a common neutral.

In fact, not doing bonding and using a common neutral takes the inverter out of the bonding game and lets the local rules do what is appropriate.
 
Shimmy said:
The problem the manufacturers are trying to address is loss of the neutral-ground bond upstream of their equipment. It is easier and cheaper (especially when talking about different international neutral bonding schemes) to consistently require a dedicated neutral (and not use an incoming neutral).
Click to expand...
You are exactly right.
These AIO's are designed by engineers in a country where the N/G bond comes from the grid. And they don’t understand the grounding system of North America.
 
l00semarble said:
But isn’t an inverter going to always be considered a “separately derived” source?
Click to expand...
In NEC terminology, the primary thing that makes something a separately derived source is if it has it's own N-G bond. So if an inverter injects a NG bond, it is a separately derived source. If an inverter does not inject an NG bond it is NOT a separately derived source.
 
FilterGuy said:
In NEC terminology, the primary thing that makes something a separately derived source is if it has it's own N-G bond. So if an inverter injects a NG bond, it is a separately derived source. If an inverter does not inject an NG bond it is NOT a separately derived source.
Click to expand...
Close, but not exactly.
What makes it a separately derived system is if the common neutral is separated and broken during the transfer.
 
timselectric said:
You are exactly right.
These AIO's are designed by engineers in a country where the N/G bond comes from the grid. And they don’t understand the grounding system of North America.
Click to expand...
You may be correct, but that still does not answer the original question: What is dangerous about using these in a common neutral situation if the bonding screws are removed?

Maybe I am just dense, but I still don't see why using these with common neutral and no bonding screws is dangerous. I have a suspicion that if we could actually have a conversation with the electrical engineers that designed the units, we would find that the inverter can be used with common neutral if the bonding screws are removed. However, between the engineer and the customer there are a LOT of people that don't understand this stuff (this includes the US-based support engineers at the company and at the distributors.). I know this is a problem because of the number of times I have talked to the US-based support engineers that are clueless about even the most basic aspects of the inverters let alone a subject as detailed as this one. The unfortunate result is that by the time it gets to the customer, the message is that it is dangerous.
 
You must log in or register to reply here.
Back
Top